Poulson Microarchitecture Article

Article: Poulson: The Future of Itanium Servers
By: David Kanter (dkanter.delete@this.realworldtech.com), May 22, 2011 10:47 pm
Room: Moderated Discussions
As of late, I've been mostly writing smaller and more digestible articles. However, I still intend to continue our tradition of in-depth microarchitecture analysis. This week's article is a analysis of Poulson and Tukwila, that builds on the disclosures at ISSCC to present the most comprehensive picture of the future Itanium architecture.

Some might question the significance of Itanium in light of the reduced ecosystem support and Oracle's transparent attempt to sabotage a rival architecture. However, Poulson promises to be a very interesting design that moves from a simple static VLIW to a more conventional pipeline.

Poulson has a totally new core with dynamic scheduling, fine-grained multithreading and a shared L3 cache. The net result is a vastly more efficient microprocessor that should achieve 2.5-2.8X higher performance and power high-end servers for the next 10 years.

In this article, we go through the microarchitecture of Poulson including the core pipeline, system interfaces and physical implementation.

http://www.realworldtech.com/page.cfm?ArticleID=RWT051811113343


Enjoy, and as always, comments and discussion are encouraged.

David
 Next Post in Thread >
TopicPosted ByDate
Poulson Microarchitecture ArticleDavid Kanter05/22/11 10:47 PM
  Itanium achieved "incredibly high single threaded performance"?Mark Roulo05/23/11 06:46 AM
    Itanium achieved "incredibly high single threaded performance"?Daniel Bizo05/23/11 07:08 AM
      Itanium achieved "incredibly high single threaded performance"?David Kanter05/23/11 07:44 AM
        Itanium achieved "incredibly high single threaded performance"?Michael S05/23/11 12:20 PM
          Itanium achieved "incredibly high single threaded performance"?Michael S05/23/11 02:30 PM
        not reallyRichard Cownie05/23/11 02:37 PM
          you are just spouting anti-IPF propagandaDaniel Bizo05/23/11 03:10 PM
            you are just spouting anti-IPF propagandaRichard Cownie05/23/11 04:13 PM
              you are just spouting anti-IPF propagandaLlew05/23/11 11:04 PM
                you are just spouting anti-IPF propagandaLinus Torvalds05/24/11 06:58 AM
                  you are just spouting anti-IPF propagandaLlew05/25/11 12:28 AM
                    you are just spouting anti-IPF propagandaBrett05/25/11 06:33 AM
                      you are just spouting anti-IPF propagandasomeone05/25/11 07:22 AM
                        you are just spouting anti-IPF propagandaRakesh Malik05/25/11 08:25 AM
                          you are just spouting anti-IPF propagandaRichard Cownie05/25/11 08:56 AM
                            you are just spouting anti-IPF propagandaRakesh Malik05/25/11 09:06 AM
                              you are just spouting anti-IPF propagandaRichard Cownie05/25/11 10:13 AM
                                you are just spouting anti-IPF propagandaRakesh Malik05/25/11 10:22 AM
                                  you are just spouting anti-IPF propagandaMichael S05/26/11 02:43 AM
                                    POWERx and ItaniumDavid Kanter05/26/11 03:53 AM
                                    you are just spouting anti-IPF propagandasomeone05/26/11 07:22 AM
                                      you are just spouting anti-IPF propagandaDavid Kanter05/26/11 07:40 AM
                                        you are just spouting anti-IPF propagandasomeone05/26/11 08:11 AM
                                          you are just spouting anti-IPF propagandaMichael S05/26/11 09:43 AM
                                            Bandwidth in LLCDavid Kanter05/26/11 02:47 PM
                                              Bandwidth in LLCIntelUser200005/26/11 11:47 PM
                                              Bandwidth in LLCMichael S05/27/11 12:45 AM
                                                Bandwidth in LLCDavid Kanter05/27/11 09:03 AM
                                you are just spouting anti-IPF propagandajrl05/25/11 05:58 PM
                                  you are just spouting anti-IPF propagandaMichael S05/26/11 02:28 AM
                                    obsess much?someone05/26/11 07:14 AM
                                      obsess much?Michael S05/26/11 08:56 AM
                                        obsess much?someone05/26/11 11:00 AM
                                    you are just spouting anti-IPF propagandajrl05/27/11 06:46 PM
                                      you are just spouting anti-IPF propagandaanonymous05/27/11 10:03 PM
                        you are just spouting anti-IPF propagandaBrett05/25/11 10:56 AM
                          you are just spouting anti-IPF propagandaDavid Kanter05/25/11 01:00 PM
                            you are just spouting anti-IPF propagandaRichard Cownie05/25/11 02:23 PM
                              Large cachesDavid Kanter05/25/11 05:22 PM
                                Large cachesRichard Cownie05/25/11 06:18 PM
                                  Large cachesanon05/26/11 04:03 PM
                                    Large cachesRichard Cownie05/26/11 04:50 PM
                                      Large cachesanon05/26/11 05:36 PM
                                        Large cachesRichard Cownie05/26/11 06:46 PM
                                          Large cachesanon05/26/11 10:43 PM
                                            Large cachesRichard Cownie05/27/11 01:24 AM
                                              Large cachesDavid Kanter05/27/11 09:20 AM
                                                Large cachesRichard Cownie05/27/11 09:51 AM
                                                  Large cachesAaron Spink05/27/11 10:18 AM
                                                    Large cachesRichard Cownie05/27/11 10:43 AM
                                                      Large cachessomeone05/27/11 12:00 PM
                                                        Large cachesRichard Cownie05/27/11 03:59 PM
                                                          WorkstationsDavid Kanter05/27/11 07:25 PM
                                                            WorkstationsRichard Cownie05/28/11 04:21 AM
                                                              WorkstationsMichael S05/29/11 02:25 AM
                                                          Large cachesAaron Spink05/27/11 09:11 PM
                                                    Large cachesCarlie Coats05/27/11 12:33 PM
                                                      Large cachesAaron Spink05/27/11 01:28 PM
                                              Large cachesanon05/27/11 04:39 PM
                                          Large cachesAaron Spink05/27/11 10:11 AM
                                            Large cachesLinus Torvalds05/27/11 12:38 PM
                                              Large cachesAaron Spink05/27/11 01:33 PM
                                                Large cachesLinus Torvalds05/27/11 03:26 PM
                                                  Miss distanceDavid Kanter05/27/11 07:01 PM
                                                    Miss distanceLlew05/27/11 10:27 PM
                                                      Miss distanceanon05/27/11 11:00 PM
                                                      Miss distancerwessel05/28/11 12:01 AM
                                                        Miss distanceLlew05/28/11 01:14 AM
                                                    Miss distanceLinus Torvalds05/28/11 06:55 AM
                                                      Miss distancea reader05/28/11 02:50 PM
                                                        Miss distanceLinus Torvalds05/28/11 03:08 PM
                                                          Miss distancea reader05/29/11 06:32 AM
                                                    Miss distanceajensen05/28/11 11:14 PM
                                                      Miss distanceLlew05/29/11 01:01 AM
                                                      Miss distancebakaneko05/29/11 03:20 AM
                                                  Run-ahead gives MLP for IOES. Rao05/28/11 04:45 PM
                                                    Run-ahead gives MLP for IOELinus Torvalds05/28/11 07:53 PM
                                                      Run-ahead gives MLP for IOES. Rao05/28/11 08:31 PM
                                                        Run-ahead gives MLP for IOELlew05/28/11 11:03 PM
                                                          Run-ahead gives MLP for IOEAaron Spink05/29/11 02:03 AM
                                                            Run-ahead gives MLP for IOEa reader05/29/11 06:27 AM
                                        Large cachesDavid Kanter05/26/11 08:58 PM
                                          Large cachesanon05/26/11 10:41 PM
                                            Large cachesRichard Cownie05/27/11 04:46 AM
                                              Large cachesanon05/27/11 09:07 PM
                                Large cachesBrett05/26/11 08:28 AM
                                  Large cachesDan Downs05/26/11 12:01 PM
                                    Large cachesDavid Kanter05/26/11 02:52 PM
                                      Large cachesMichael S05/27/11 01:01 AM
                                    Large cachesBrett05/26/11 07:38 PM
                                      Large cachesDan Downs05/27/11 09:59 AM
                                        Large cachesBrett05/28/11 09:45 AM
                            you are just spouting anti-IPF propagandampx05/26/11 11:57 AM
                        you are just spouting anti-IPF propagandaMichael S05/26/11 03:01 AM
                          you are just spouting anti-IPF propagandaanon05/26/11 04:21 PM
                      you are just spouting anti-IPF propagandaanon05/25/11 08:13 AM
                  you are just spouting anti-IPF propagandaDavid Kanter05/25/11 11:45 PM
                    you are just spouting anti-IPF propagandasomeone05/26/11 07:55 AM
                      Poulson power and perfDavid Kanter05/26/11 08:10 AM
                        Poulson power and perfsomeone05/26/11 08:40 AM
                        Poulson power and perfanon05/26/11 08:31 PM
                          Poulson power and perfDavid Kanter05/26/11 09:12 PM
                            Poulson power and perfanon05/28/11 07:24 AM
                              Poulson power and perfDavid Kanter05/28/11 09:28 AM
              you are just spouting anti-IPF propagandaDaniel Bizo05/23/11 11:30 PM
                you are just spouting anti-IPF propagandaRichard Cownie05/24/11 01:28 AM
                you are just spouting anti-IPF propagandampx05/25/11 11:03 AM
                  you are just spouting anti-IPF propagandaMichael S05/26/11 02:06 AM
                    you are just spouting anti-IPF propagandampx05/26/11 11:27 AM
            highest performing xx nm CPUsMichael S05/24/11 07:55 AM
      Itanium achieved "incredibly high single threaded performance"?Max05/23/11 09:07 AM
        Itanium achieved "incredibly high single threaded performance"?Daniel Bizo05/23/11 12:59 PM
    Itanium achieved "incredibly high single threaded performance"?David Kanter05/23/11 07:14 AM
      Itanium achieved "incredibly high single threaded performance"?Michael S05/23/11 12:17 PM
        Itanium achieved "incredibly high single threaded performance"?David Kanter05/23/11 04:56 PM
          Itanium achieved "incredibly high single threaded performance"?Thu Nguyen05/23/11 08:09 PM
            Itanium achieved "incredibly high single threaded performance"?David Kanter05/23/11 09:11 PM
              Itanium achieved "incredibly high single threaded performance"?Michael S05/24/11 02:08 AM
                Itanium achieved "incredibly high single threaded performance"?David Kanter05/24/11 07:36 AM
                  Itanium achieved "incredibly high single threaded performance"?Michael S05/24/11 08:17 AM
                  Itanium achieved "incredibly high single threaded performance"?Thu Nguyen05/24/11 12:45 PM
                    POWER7 vs. SNB SPECint_baseDavid Kanter05/24/11 08:15 PM
                      SNB SPECint linkDavid Kanter05/24/11 08:17 PM
                      Lies, damn lies & SPEC resultsPassing by05/26/11 06:06 AM
                        Lies, damn lies & SPEC resultsDavid Kanter05/26/11 07:17 AM
                          Lies, damn lies & SPEC resultsMichael S05/26/11 09:57 AM
                          Lies, damn lies & SPEC resultsanon05/26/11 04:44 PM
                        Lies, damn lies & SPEC resultsMichael S05/26/11 10:04 AM
                          Lies, damn lies & SPEC resultsLinus Torvalds05/26/11 10:53 AM
                            Lies, damn lies & SPEC resultsJouni Osmala05/26/11 11:42 PM
                              mem usage distributionMichael S05/27/11 12:38 AM
                                mem usage distributionJouni Osmala05/27/11 03:25 AM
                              Lies, damn lies & SPEC resultsLinus Torvalds05/27/11 05:44 AM
                                Lies, damn lies & SPEC resultsJouni Osmala05/28/11 04:19 AM
                                  Lies, damn lies & SPEC resultsLinus Torvalds05/28/11 07:29 AM
                                    Using larger translationsPaul A. Clayton05/28/11 11:17 AM
                                      Using larger translationsLinus Torvalds05/28/11 11:58 AM
                                        Using larger translationsanon05/28/11 03:37 PM
                                        Large pages still seem attractivePaul A. Clayton05/28/11 07:47 PM
                                          Large pages still seem attractiveLinus Torvalds05/28/11 09:02 PM
                                    Lies, damn lies & SPEC resultsDoug Siebert05/28/11 11:35 AM
                                      Lies, damn lies & SPEC resultsLinus Torvalds05/28/11 12:14 PM
                                        Lies, damn lies & SPEC resultsLinus Torvalds05/28/11 12:25 PM
                                        Lies, damn lies & SPEC resultsmpx05/29/11 07:12 AM
                                          Lies, damn lies & SPEC resultsLinus Torvalds05/30/11 05:04 AM
                                            Lies, damn lies & SPEC resultsgallier205/30/11 06:23 AM
                                            Lies, damn lies & SPEC resultsmpx05/30/11 12:49 PM
                                        Lies, damn lies & SPEC resultsDoug Siebert05/29/11 10:30 AM
                                      Lies, damn lies & SPEC resultsanon05/28/11 02:54 PM
                                        Lies, damn lies & SPEC resultsnone05/28/11 03:58 PM
                                          How's that for compatibility?David Kanter05/28/11 04:14 PM
                                            How's that for compatibility?none05/28/11 04:38 PM
                                              How's that for compatibility?anon05/28/11 07:00 PM
                                                How's that for compatibility?Jouni Osmala05/28/11 10:32 PM
                                                  Larger caches with smaller pagesPaul A. Clayton05/29/11 07:40 PM
                                                How's that for compatibility?Wilco05/30/11 09:18 AM
                                                  How's that for compatibility?anon05/30/11 09:20 PM
                                                    How's that for compatibility?Wilco06/01/11 02:47 AM
                                                      How's that for compatibility?anon06/01/11 04:12 AM
                                                        How's that for compatibility?Michael S06/02/11 12:41 AM
                                                          How's that for compatibility?Wilco06/02/11 02:01 AM
                                                            How's that for compatibility?none06/02/11 08:03 AM
                                      Lies, damn lies & SPEC resultsanon05/28/11 05:02 PM
                                        Lies, damn lies & SPEC resultsDavid Kanter05/28/11 06:21 PM
                                        Lies, damn lies & SPEC resultsS. Rao05/28/11 06:34 PM
                                          Lies, damn lies & SPEC resultsanon05/28/11 10:47 PM
                                            Lies, damn lies & SPEC resultsS. Rao05/29/11 12:09 AM
                                              Lies, damn lies & SPEC resultsMichael S05/29/11 01:59 AM
                                                Lies, damn lies & SPEC resultsanon05/29/11 02:45 AM
                                                  Lies, damn lies & SPEC resultsanon05/29/11 03:18 AM
                                            Lies, damn lies & SPEC resultsJouni Osmala05/29/11 12:49 AM
                                              Lies, damn lies & SPEC resultsanon05/29/11 02:32 AM
                                                Lies, damn lies & SPEC resultsJouni Osmala05/29/11 03:19 AM
                                                  Lies, damn lies & SPEC resultsMichael S05/29/11 08:05 AM
                                                    Lies, damn lies & SPEC resultsJouni Osmala05/29/11 08:40 AM
                                                      Lies, damn lies & SPEC resultsMichael S05/29/11 09:25 AM
                                                  Lies, damn lies & SPEC resultsanon05/29/11 10:50 AM
                                              Lies, damn lies & SPEC resultsmpx05/30/11 02:22 AM
                                      Lies, damn lies & SPEC resultsGabriele Svelto05/29/11 01:09 AM
                          Lies, damn lies & SPEC resultsanon05/26/11 04:58 PM
                  Itanium achieved "incredibly high single threaded performance"?anon05/24/11 09:25 PM
    Itanium achieved "incredibly high single threaded performance"?someone05/24/11 05:40 AM
      Itanium achieved "incredibly high single threaded performance"?none05/24/11 07:08 AM
        Itanium achieved "incredibly high single threaded performance"?someone05/24/11 08:35 AM
      Itanium achieved "incredibly high single threaded performance"?Megol05/24/11 07:10 AM
      Itanium achieved "incredibly high single threaded performance"?Seni05/24/11 07:43 AM
        Itanium achieved "incredibly high single threaded performance"?someone05/24/11 08:38 AM
      Itanium achieved "incredibly high single threaded performance"?Mark Roulo05/24/11 01:41 PM
        P4 180 nm with L3anonymous05/25/11 07:16 AM
      Itanium achieved "incredibly high single threaded performance"?TA152H05/25/11 08:58 AM
  Thanks!Wouter Tinus05/23/11 10:21 AM
  Poulson Microarchitecture Articleajensen05/23/11 11:16 AM
    Poulson Microarchitecture ArticleDavid Kanter05/23/11 04:48 PM
      Poulson control queueajensen05/23/11 08:48 PM
        Poulson control queueDavid Kanter05/25/11 11:22 PM
  Some nitpicks and comments/questionsPaul A. Clayton05/23/11 11:22 AM
    Wide-Word [was: Some nitpicks and comments/questions]anon05/23/11 07:33 PM
    Some nitpicks and comments/questionsDavid Kanter05/24/11 08:36 AM
      Little technically keeps Xeon from true high-endPaul A. Clayton05/24/11 02:56 PM
        Little technically keeps Xeon from true high-endDavid Hess05/24/11 06:16 PM
          ChipKill implementation requirementsPaul A. Clayton05/25/11 07:38 AM
            ChipKill implementation requirementsDavid Hess05/25/11 04:10 PM
  spec int base referenceajensen05/24/11 12:08 PM
    spec int base referenceMichael S05/24/11 01:08 PM
    spec int base referenceMichael S05/24/11 02:02 PM
      spec int base referenceajensen05/24/11 09:38 PM
        spec int base referenceMichael S05/25/11 12:44 AM
Reply to this Topic
Name:
Email:
Topic:
Body: No Text
How do you spell blue?